Chemical principles of single-molecule electronics - …nuckolls.chem.columbia.edu/system/files/175/original/16… · · 2017-01-29Chemical principles of single-molecule electronics

Over the past century, experimental physicists and engineers have developed sophisticated methods to create semiconducting silicon-based devices, such as diodes, transistors and memory elements, of increas-ingly small dimensions. Meanwhile, chemists have acquired a detailed understanding of the relationship between the chemical structure and the electronic properties of a molecule through reaction chemistry, advanced physical chemistry and theoretical methods. These multidisciplinary efforts have converged in the field of single-molecule electronics (SMEs), in which the ultimate goal is to use molecules as active elements in electronic circuitry1,2. A rich knowledge base detail-ing the electronic properties of molecules already exists in the context of chemical reactivity; reimagining such properties in the framework of SMEs might then inspire tremendous advancements in thefield.

Powerful methods have been developed to charac-terize and manipulate the conductance properties of single molecules36. Molecular conductance has been measured using techniques based on scanning tunnel-ling microscopy (STM)7, mechanically controlled break junctions811, STM break junctions12,13, conductive atomic force microscopy14, electromigration15, nanoparticle arrays16,17 and other approaches1821. The sophistication of these techniques provides the opportunity for chemists to collaborate with physicists and engineers to incorporate well-understood chemical principles into the study of structureconductivity relationships in molecularwires.

Reviews on SMEs are usually written from the per-spective of those who build devices and measure their properties the link between chemistry and SME devices is rarely the focus. Filling this gap will address the questions that arise from the expanding structural diversity of molecular wires in the SME literature. For example, how will the single-molecule conduct-ance paradigm, which was first developed for simple molecular structures, shift as molecular wires grow in structural complexity and diversity? Which lessons can be drawn from reaction chemistry to guide the design of molecular electronics? How can chemical expertise be used to engineer new functions into single-molecule wires?

In this Review, we integrate the languages of chemis-try and device physics to explain the chemical concepts that underlie single-molecule conductance. We discuss the structureproperty relationships of single-molecule junctions by deconstructing the junction into three distinct components: the anchor, the electrode and the bridge (FIG.1). We survey the modularity of each component and describe how tuning the structure of each part affects the charge-transport properties of the junction, primarily in the context of break-junction experiments. Finally, we examine emerging areas in SME research, such as single-molecule conductance switches and quantum interference (QI), and discuss how these are fundamentally related to well-established chemical principles.

1Department of Chemistry, Columbia University.2Department of Physics and Applied Math, Columbia University, New York, New York 10027, USA.

Correspondence to M.L.S., L.V.and [email protected];[email protected];[email protected]

Arricle number: 16002 doi:10.1038/natrevmats.2016.2Published online 23 Feb 2016

Chemical principles of single-molecule electronicsTimothy A.Su1, Madhav Neupane1, Michael L.Steigerwald1, Latha Venkataraman1,2 and Colin Nuckolls1

Abstract | The field of single-molecule electronics harnesses expertise from engineering, physics and chemistry to realize circuit elements at the limit of miniaturization; it is a subfield of nanoelectronics in which the electronic components are single molecules. In this Review, we survey the field from a chemical perspective and discuss the structureproperty relationships of the three components that form a single-molecule junction: the anchor, the electrode and the molecular bridge. The spatial orientation and electronic coupling between each component profoundly affect the conductance properties and functions of the single-molecule device. We describe the design principles of the anchor group, the influence of the electronic configuration of the electrode and the effect of manipulating the structure of the molecular backbone and of its substituent groups. We discuss single-molecule conductance switches as well as the phenomenon of quantum interference and then trace their fundamental roots back to chemical principles.

NATURE REVIEWS | MATERIALS VOLUME 1 | MARCH 2016 | 1

REVIEWS

2016 Macmillan Publishers Limited. All rights reserved

mailto:mailto:mls2064%40columbia.edu?subject=mailto:mailto:lv2117%40columbia.edu?subject=mailto:mailto:cn37%40columbia.edu?subject=

Anchor groupThe anchor group (also known as the linker or contact group) connects the molecular wire to the electrodes both mechanically and electronically. Usually, a single anchoring group terminates each end of the molecule to form the metalmoleculemetal junction; however, including more anchoring units along the molecular bridge can offer additional handles for tuning con-ductance, depending on which two anchors form the most conductive pathway2227. Anchoring groups typi-cally bind to electrodes either through donoracceptor (dative) interactions or through covalent bonding. Prototypical anchor groups for each type of electrodelinker interaction are shown in FIG.2a. Because gold is the most used electrode material in SME studies, we focus primarily on the interaction of anchor groups with gold electrodes.

Dative interactions involve the electron donation from a donor or a lone pair donor to a Lewis acidic Au atom. Common donors include fullerenes28,29 and other -conjugated hydrocarbons19,26,30,31. Many lone pair anchoring groups are common -donor ligands familiar from coordination chemistry32. Dative con-tacts, such as amines, are advantageous because they bind selectively to undercoordinated adatoms on the electrode surface; this narrows the conductance distri-bution because it limits the Aulinker contact geom-etry33. Covalent contacts between the metal and the molecule result from covalent bonding between mole-cular radicals and metallic electrode surfaces. Covalent contacts are valuable because they are physically robust linkages that strongly couple the electronics of the molecule and themetal.

Conductance depends not only on the class of the anchoring group but also on the nature of the inter-action between the anchor group and the other com-ponents of the junction that is, the bridge and the electrode. In the following sections, we explore how the spatial overlap between the orbitals of these three components affects the charge-transport properties of the junction.

Before we continue, we must clarify the meaning of the conductance values that are discussed. There is significant measurement-to-measurement variability in single-molecule experiments that is due to fluctuations in the molecular conformation, the electrodeanchor contact geometry and the electrode surface geometry. To account for this variability and to better under-stand the nature of conductance in a single molecule,

researchers analyse hundreds to thousands of measure-ment traces together by compiling them into conduct-ance histograms to obtain a distribution of all measured conductance values. The conductance values that we report here refer to the conductance peak values from such histograms reported in units of G0. G0 is the con-ductance quantum and is defined as G0 = 2e2/h = 77.5 S, where e is the charge of an electron and h is Plancks constant; it is the preferred unit used to describe the conductance between metal point contacts as well as molecular conductance.

Effect of anchorbridge orbital overlap on conductance. The anchor group often dictates whether the mole cular wire transports holes (highest occupied molecular orbital (HOMO)-dominated conductance) or elec-trons (lowest unoccupied molecular orbital (LUMO)-dominated conductance). The dominant conducting molecular orbital is typically the orbital that is closest in energy to the electrode Fermi level, EF. Conductance depends on the energy offset, E, between EF and the conducting orbital, and on the strength of the elec-trodemolecule hybridization, (BOX1). The nature of the conducting orbital can be determined experimen-tally through thermopower measurements34,35 or com-putationally by transmission calculations36. However, in simple structures, we can predict the nature of charge carriers from basic chemical principles by considering the nature of the molecular backbone and the geome-try of the lone pair relative to the conjugated orbitals of the molecule (FIG.2b). To illustrate this method, we use a phenyl ring terminated at the para positions by the dative linker groups from FIG.2a. This analysis can be applied to other basic aromatic wires as well. For ben-zene rings terminated with linkers such as SR, NH2, PR2 or SeR at the 1,4-positions, the lone pair orbitals are included in the HOMO because they are coplanar with and energetically destabilized by the filled -con-jugated bridge orbitals. Thus, owing to the alignment of the Aulone pair bonds with the system of the bridge, conductance occurs strongly through the HOMO when such anchor groups areused.

By contrast, conductance in phenyl rings terminated by pyridine37, isocyanide38 and cyanide39 linkers occurs primarily through the LUMO. For these anchor groups, the lone pair lies in the plane of the molecule, rigidly orthogonal to the channel of the wire. Conduction through the lone pair orbitals is weak because the car-bon sp2 orbitals are poorly conjugated40. Moreover, the lone pair orbital is generally quite low in energy because it is part of the system; thus, transport through this orbital has a marginal contribution to conductance (E is large). Just as importantly, the elec-tron-withdrawing nature of these linker groups facili-tates LUMO-dominated conductance by lowering the energy of the *-antibonding orbitals towards the EF (and the HOMO-conducting -bonding orbitals away from the EF). The conductance is then controlled by the coupling between the electrodes and the *-antibond-ing orbitals of the LUMO. When molecular bridges are very electron-deficient structures, such as thiophene

Electrode Anchor ElectrodeAnchorBridge

I

Nature Reviews | MaterialsFigure 1 | A schematic of a single-molecule junction with electrode, anchor and bridge components. The bridge unit can be further deconstructed into backbone (blue block) and substituent (red circles) subunits. I, current.

R E V I E W S

2 | MARCH 2016 | VOLUME 1 www.nature.com/natrevmats


dioxides41 and porphyrins42, in which the HOMO and LUMO energies are substantially lowered, conductance through the LUMO can dominate regardless of the type of linkerused.

HOMO and LUMO conduction can also be understood from the perspective of coordination chemistry and of the different modes of interaction between ligands and transition metals. Hole trans-port or HOMO-dominated conduction occurs when the metalmolecule bond, formed using the -donor orbital of the molecule (which in most cases is the HOMO of the isolated molecule) and the -accepting orbital of the metal, can become coplanar with the con-jugated bridge orbitals. This coplanarity ensures that the metalmolecule bond, the gateway, can mix with the delocalized, conjugated bridge orbitals and estab-lish the conductivity path. Conversely, if the metal- to-molecule bond cannot mix with the conjugated pathway, the charge carriers cannot use the molecular HOMO; it is geometrically unavailable and energet-ically distant from the EF. However, it is well known from coordination and organometallic chemistry that although ligand-to-metal donation usually dominates metalligand bonding, it is often supplemented by dmetalto*ligand back-bonding, in which occupied orbit-als of the metal mix with unoccupied orbitals of the

ligand. In the case of molecular conduction, when the -donation (HOMO-transporting) pathway is un avail-able, this * back-bonding (LUMO-transporting) path-way may be available if the * orbital, into which the metal back- donates, is a conjugated orbital that spans the molecule and is connected to both electrodes.

The stereochemistry of the Auanchor bond with respect to the conductive channel of the molecular bridge determines the strength of the electronic cou-pling between the metal and the molecule; manipulating this stereochemistry is a powerful handle for controlling conductance. The charge flow between the electrodes increases if the Auanchor bond is aligned with the con-jugated orbitals of the molecule. The position of sulfur lone pairs can be locked into alignment with the molec-ular backbone using a dihydrobenzothiophene (BT) thioether linker43. The frustrated rotation of the S lone pair results in increased conductance and in a sharper conductance peak compared with analogous aromatic wires with thiomethyl linkers that can freely rotate. The BT linker has been incorporated into several dif-ferent molecular wires to strengthen the anchorbridge coupling27,4446.

Poor coupling between the electrode and the mole cule can also be a desirable quality. Electrode molecule coupling can be disrupted by inserting

Nature Reviews | Materials

a

b

H2N-Ph-NH

2 (HOMO) NC-Ph-CN (HOMO-2) NC-Ph-CN (LUMO)

NH N2

R P2

C N

RS

RSe

Dative anchors Covalent anchors

HSS

AuAu

R Sn3 AuAu

Me Si3 AuAuTBAF

Lone pair donors donors

AuAu+ e

N N+

N C

Figure 2 | Anchor group archetypes and the nature of charge carriers for common dative anchors. a|Molecular structures of common anchors. Dative anchors can be classified as donating or lone pair donating. For lone pair donors, the anchors shown in the left and in the right columns impart lowest unoccupied molecular orbital (LUMO)- and highest occupied molecular orbital (HOMO)-dominated conductance, respectively, in simple -conjugated systems. Covalent anchors commonly used to generate direct AuS and AuC contacts are shown in the last column. These contacts can be generated from thiol oxidation, AuSn transmetalation, fluoride-initiated desilylation and diazonium electroreduction reactions. b|The highest energy molecular orbital surfaces that feature strong lone pair character are depicted for 1,4-diaminobenzene (left panel) and 1,4-dicyanobenzene (middle panel) (B3LYP/6-31G**). In 1,4-diaminobenzene, the N-centred lone pairs occupy p orbitals that, like the benzene orbitals, are perpendicular to the plane of the ring. In 1,4-dicyanobenzene, these N-centred lone pair orbitals are orthogonal to the channel of the benzene ring; conductance in this system is dominated instead by transport through the LUMO (right panel).

R E V I E W S



methylene (CH2) units between the S atom and the phenyl ring. The methylene spacers allow gating effects, such as Coulomb blockade, to occur in three-electrode systems because they decouple the molecule from the source and drain the electrodes47 (FIG.3a). However, the reduction in moleculeelectrode coupling has been shown to decrease the junction conductance by three orders of magnitude in -conjugated molecules47. Interesting functions can be implemented into molec-ular electronic devices by synthetically engineering anchor groups with both strong and weak electrodemolecule coupling character. A bulky SPh anchor group can be used to misalign the SAu bond with the molecular bridge, decreasing the moleculeelectrode coupling45. This enables the creation of a single-mole-cule rectifier, whereby the molecule is strongly coupled to the electrode by a covalent AuC bond at one end and weakly coupled by a AuSPhR bond at the other end. A class of oligosilanes and oligogermanes that switch between different conductance values depending on the strength of the coupling between the electrode

and the molecule has recently been described48,49. Junction elongation stretches the terminal ends of the molecule into a geometry (orthoortho (OO)) that is optimal for conductance and electrodemolecule cou-pling; junction compression relaxes the molecule into dihedral geometries (antianti (AA) or orthoanti (OA)) with diminished electrodemolecule coupling and lower conductance (FIG.3b).

In situ chemical reactions to produce covalent contacts. AuS linkages formed from the reduction of thiol (SH) linkers on Au surfaces are the most widely studied form of covalent contact. Strong AuS linkages ena-ble the molecular junction to withstand harsh external conditions such as mechanical stress50 and high-bias voltages51. Thiols can oxidize to disulfides rather easily under ambient conditions; this is problematic because dithiolated molecular wires can polymerize and form insoluble polydisulfides52. Although SS bonds can be reduced on gold surfaces53, there is no guarantee that the reduction will be exhaustive and that only the


E

E

EF

EF

EF

HOMO

LUMO

/2 /2

E

Box 1 | Principles of charge transport in single-molecule junctions

In the limit of coherent transport, the conductance of a molecular junction is related to the probability of electron transmission between the electrodes through the Landauer formula143,144. Often, this transmission probability strongly depends on the extent of the hybridization of a single molecular orbital with the continuous band of energy levels on the metal electrode. In these cases, transmission depends primarily on the energetic alignment (E) of this orbital with respect to the metal Fermi level (EF) and on the broadening () of the orbital due to hybridization, which gives electrons on this level a finite lifetime ~/. In the figure, the red Lorentzian curves (lowest unoccupied molecular orbital (LUMO)) and blue Lorentzian curves (highest occupied molecular orbital (HOMO)) depict a typical transmission function plotted against energy. The transmission function shown here describes HOMO-dominated conductance because the HOMO transmission peak is closest in energy to and intersects the EF; a similar picture for LUMO-dominated conductance would show the LUMO transmission peak intersecting the EF.

The parameters and E depend on intrinsic molecular and metallic characteristics, including the nature of the molecular orbital, the moleculemetal bond character (covalent or donoracceptor), the electronic density of states of the metal and the charged interactions at the metalmolecule interface. Orbitals that have density on the metal-binding anchor groups generally result in a broadened transmission resonance (large ), whereas those that localize electron density on the molecular backbone show very narrow resonances (small). The length of the molecule also effects the width of the resonance; small molecules, such as H2 and short -conjugated wires, are strongly coupled because their orbitals effectively hybridize with the metal, which results in a large. For longer molecular wires, is smaller because a larger proportion of the hybridized orbital is localized on the molecular backbone. In oligomeric materials, decreases with increasing n more quickly if the conjugation or coupling strength () between oligomer units is weak. This is related to the reason why the length-dependent conductance decay parameter () is small for conjugated repeat units but large for non-conjugated repeat units. The lower panel of the figure schematically represents a tight-binding model that assigns a single level () to each repeat unit and the parameter that describes the electronic coupling between them. The coupling of these units with the electrodes is described by /2.

R E V I E W S



monomers will contribute to conduction measure-ments. A common approach for increasing the ambi-ent stability of thiol-based wires is to functionalize the thiols with thioacetate-protecting groups that can cleave on the electrode surface to form covalent AuS contacts52,54. Thiol-based junctions tend to show broad conductance features owing to the large variability in the anchorelectrode contact geometry. Several groups have studied the effect of binding geometry on conduct-ance5559, but understanding and thus gaining control over the variability of contact geometry is an ongoing challenge in the field42,6062.

AuC contacts are particularly promising covalent anchors because they give well-defined conductance peaks owing to the selective binding to undercoordi-nated gold63. Furthermore, molecular wires with AuC contacts generally demonstrate higher conductance peak values than structurally analogous wires with AuS contacts. For example, the AubenzenedithiolAu junction shows a broad conductance, with reported conductance peak values ranging from 102 to 104 G0 (REFS11,6466). By contrast, the structurally similar AuxylylAu junction conducts at 0.9 G0 (REF.67). The difference in conductance arises from two factors. First,


a

dc

e

Cur

rent

(pA

)

Log(

GX/G

H)/

N

Ape

x de

nsit

y of

sta

tes

(10

2 eV

1)

Cur

rent

(A

)

Diff

erential conductance (1/G

)

Diff

erential conductance (1/M

)Bias (mV)Bias (mV)

Energy EF (eV)

para (benzoic acid)

0.0 0.5 1.00.51.0 0.4 0.6 0.80.00.2 0.2

5000 1,0005001,0005000 1,0005001,00080

60

40

20

0

20

40

60

2.4

1.6

0.8

0

0.8

1.6

2.4

0.04

0.02

0.00

0.02

0.04

0

1

2

3

4

5

6

1.2

0.9

0.6

0.3

0

0.3

0.6

0.9

0

1

2

3

4

5

6

Ag dyz

Ag py

Au dyz

Au py

Twist angle

Conductance

HC

2

SAu

SAu

31

2

4 5

6

127

9

11

810

b

SiSiMe2

H HAu

S Me

SiSiMe2

H H

Au

SMe

SiSiMe2

H H

Au

S Me

Au

SiSiMe2

H HMe

S Au

SiSiMe2

H HMe

S

Au

SiSiMe2

H H

S

Me

Com

pres

sion

Elongation

Low G state

AA

OA

High G state

OO

S SiSi

SiSi S

Au

Au

CurrentDifferentialconductance

Figure 3 | Tuning the structure of the anchor, electrode and bridge to modulate charge-transport properties in single-molecule junctions. a|Current and differential conductance plotted against source-drain voltage bias at a constant 1V gate voltage for an oligo(pphenylene vinylene) molecule with (left panel) and without (right panel) methylene insertions adjacent to the thiol. The tunnelling barrier introduced by the methylene unit decouples the molecule from the electrodes and a Coulomb blockade is observed. b|Newman projections depicting the anti (A) and ortho (O) terminal dihedral conformations in a AuoligosilaneAu junction. Junction elongation causes the terminal electrodeanchor dihedral geometry to shift from low conducting (low G state), with weakly coupled AA and OA configurations, to high conducting (high G state), with a strongly coupled OO configuration. c|Density of states around the Fermi level (EF) for py and dyz orbitals in Ag and Au. The stronger d-orbital character of Au compared with Ag near the EF results in a stronger d * coupling for pyridine anchors and, therefore, an increased conductance. d|The Hammett plot shows the relationship between conductance and each substituents Hammett coefficient para in substituted benzene rings. e|Conductance decreases with an increasing twist angle between biphenyl rings. GH, conductance of unsubstituted molecule; GX, conductance of substituted molecule; N, number of substituents. Panel a is adapted with permission from REF.47, American Chemical Society. Panel b is from REF.48, Nature Publishing Group. Panel c is adapted with permission from REF.92, American Chemical Society. Panel d is reproduced with permission from REF.103, American Chemical Society. Panel e is from REF.13, Nature Publishing Group.

R E V I E W S



the CAu bond is more strongly coupled to the system than the SAu bond because it is much shorter in bond length. Second, E is much smaller for the AuC gate-way states that describe the covalent metalmolecule hybridization. For example, the AuS gateway state is centred at E EF = 1.4 eV for alkane dithiols59, whereas the AuC gateway state is centred at E EF = 0.8 eV for bis(trimethylstannyl)alkanes67. This difference in energy alignment with the EF contributes significantly to the difference in conductance.

Three methods have been developed for the insitu generation of direct AuC covalent contacts. The first method involves the transmetalation of CSnR3 bonds on gold surfaces to generate CAu bonds (and tin oxide by-products under ambient conditions68). This method was first used to obtain self-assembled alkane monolay-ers on gold surfaces from organotin species69, and it was later used for the insitu cleavage of terminal CSnMe3 bonds to obtain covalent Auarene and Aualkane con-tacts in single-molecule junctions67. Alkane67 and para-phenylene70 wires terminated with AuCH2R contacts demonstrate a 10- to 100-fold increase in conductance compared with the analogous bridges terminated with dative AuNH2R contacts. The applicability of this method was recently expanded to include Auacetylene contacts71. A potential shortcoming of this approach is that it both uses and produces toxic and volatile trimeth-yltin species. Furthermore, this reaction does not occur universally for all organotin molecules; for example, this manner of cleavage does not occur in perfluorinated benzene backbones70. The most important molecu-lar design rule for creating CAu contacts via CSnR3 transmetallation on gold is that the bond between the molecular bridge and the tin atom must be the most reactive of the four organotin bonds. For example, nal-kane backbones with SnBu3 end groups do not show clean conductance features in the STM break junction67, presumably because the cleavage of the four CSn bonds is not selective. By contrast, Au(CH2)nAu junctions form cleanly for trimethyltin-terminated alkane wires because of the greater stability of RH2C radicals com-pared with H3C, which enables the preferential cleav-age of the RH2CSn bond. Similarly, benzyltrimethyltin molecules cleave instantaneously at the Snbenzyl bond, even at 110 C (REF.63), whereas Snaryl bonds cleave slowly, with AuarylAu junctions appearing only after 2.5hours at room temperature67. This design strategy allows the programming of the junction that will form on SnC cleavage.

The second method for obtaining covalent AuC contacts involves a fluoride-initiated desilylation of oligo(phenylene ethynylene) wires terminated with trimethylsilyl (TMS) end groups72. Addition of tetrabutylammonium fluoride to a solution of the TMS-protected target molecules selectively cleaves the termi-nal ethynylSi bonds. This approach is inspired from a classic synthetic chemistry method that exploits the strong affinity between silicon and fluorine to unmask acetylene groups73. The applicability of this method is hindered by the NBu4+ electrolytes that participate in ionic conductance between the electrodes; these

electrolytes give rise to significant conductance noise that may cover the signal of low-conductance mole-cules. However, ionic conductance can be reduced by coating the electrodes with an insulating layer74.

The third method to produce covalent electrode anchors involves the electroreduction of diazonium salts on gold surfaces75. It was first used in the context of break-junction experiments by electrochemically reducing the terminal diazonium end groups on a biphenyl ring to generate covalent AubiphenylAu junctions76. This approach is attractive because cova-lent AuC contacts can be generated on demand by increasing the reduction potential via a gate electrode to irreversibly cleave the arylN bond. However, diazonium salts are known to be thermally unstable and, in many cases, explosive77. In particular, alkyl diazonium salts are especially unstable, which limits the range of diazonium-functionalized structures that can be easily measured in single-molecule junctions.

There are still many unsolved issues in the imple-mentation of single-molecule devices with covalent contacts. The choice of precursors for the desilylation and diazonium reduction methods has been limited thus far to those that place the AuC bonds in the plane of the molecular bridge. This is an important con-sideration because maintaining coplanarity between the metalcarbon bond and the bridge system is essential for optimizing the coupling between molecule and elec-trode. Moreover, molecular wires with AuC contacts are not particularly robust, as they tend to oligomerize insitu during break-junction experiments. This may be unavoidable under ambient conditions, as AuC bonds are inherently sensitive to oxidation and dimerization pathways. Tuning the structure of the wire to make the metalcarbon bond more stable is a possible solution to avoid device failure. Using electrode materials that form more stable electrodecarbon bonds is another possible route for enhancing the stability of the device; we discuss this topic in the following section.

ElectrodeThe electrode as a chemical reagent. Using the elec-trode as a reagent in synthetic reactions is a promising and underexplored route for the development of SME devices with desirable properties. Concepts in organo-metallic chemistry describe how the electronic struc-ture of different metals affects their chemical reactivity. Inorganic chemical principles, such as the hardsoft acidbase concept and ligand field theory, provide a general roadmap for understanding the chemical groups that can be used to functionalize electrode surfaces; soft metals that are commonly used as elec-trode materials interact strongly with soft and high-field ligands32,78, such as the ones depicted in FIG.2a. Several classic organometallic and organic reactions have already been transposed from reaction flasks to electrode surfaces. For example, the Ullmann coupling reaction, which uses the metal-mediated homocou-pling of halobenzenes to fuse aryl rings together, was discovered79 in 1901 more than a century later it was reimagined on a gold surface to synthesize graphene

R E V I E W S



nanoribbons80,81. The coupling reactions between an amine and a carboxylic acid, which are fundamental for peptide chemistry, have been used to produce covalent electrodemoleculeelectrode junctions by reacting amine-terminated molecules with carboxylate point defects in carbon-nanotube electrodes18. The ruthe-nium alkylidene chemistry familiar to olefin metathesis reactions82 has been used for functionalizing ruthe-nium electrodes with alkylidenes that are well coupled, which is relevant for charge transport, and catalytically active, so that longer wires can be grown83. As the field of SMEs develops, more examples will arise in which the reactivity profile of specific metals is exploited to functionalize electrode surfaces.

Electrode materials. Gold is the most common elec-trode material in break-junction experiments because of its inertness, which enables the measurement of single-molecule junctions with consistency and repro-ducibility under ambient conditions. Other metals have interesting electronic properties, but many of them quickly oxidize in air, creating oxide layers on the electrode surface that prevent the clean formation of metalmoleculemetal junctions. Measuring in air-free or ultrahigh vacuum conditions can help to cir-cumvent this problem but adds a significant degree of complexity to the experiment. Electrode materials that have been used for SME devices include metals such as Ag84,85, Pd86 and Pt8789, and graphitic nanostructures such as graphene19 and carbon nanotubes18.

The density of states of a metal at the EF strongly influences the conductance of the single-molecule junction. This was demonstrated in a study on isothi-ocyanate-terminated alkanes, in which the observed conductance was two- to threefold higher in Pd and Pt junctions than in Au junctions90. The metal d band possesses the appropriate symmetry to couple with the isothiocyanate orbitals near the EF; thus, the increased d character at the EF of Pd and Pt relative to Au enhances the metalmolecule d interaction91. In another study, it was found that the conductance peak value in 4,4-bi-pyridine is more than an order of magnitude lower when it is measured with Ag electrodes rather than Au electrodes. This difference in conductance arises from the weaker dyz-orbital character of the Ag density of states at the EF that results in reduced d* hybrid-ization and metalmolecule coupling92 (FIG.3c). Metals with unpaired spins can also confer interesting magnetic properties to single-molecule junctions9395. Harnessing the unique chemical and physical properties of different electrode materials will allow a greater degree of control over charge transport in single-molecule junctions and will enable the design of new SME devices.

Molecular bridgeThe idea that a molecule can function as an active com-ponent in an electrical circuit was first formulated in the context of the design of a theoretical rectifier with an asymmetric molecular bridge structure, in which charge would flow preferentially from electron-rich to elec-tron-deficient regions1. Out of the three modules of the

molecular junction, the bridge has the greatest potential for manipulation with synthetic chemistry: any chemi-cally reasonable structure can be prepared and can serve as a molecular bridge as long as it contains two anchor-ing groups. There are two distinct subcomponents in the bridge: the backbone and the substituents. The backbone is the main pathway through which charge flows such as the bonds in phenyl rings or the SiSi bonds in oligosilanes. The substituents are the chemical groups attached to the main backbone chain, and they can alter both the electronic structure and the conformation of the molecule.

values and electronic coupling in the molecular back-bone. The ability of different oligomeric backbones to transport charge can be evaluated by comparing how their conductance decays with increasing oligomer length. This is quantitatively described by their values, which are given in units of inverse length. The value is derived by plotting conductance on a semi-logarithmic scale against the molecular length (L) of the oligomer. The value is then extracted using the formula G = AeL. values depend on the coupling strength between repeat units: backbones that are strongly conjugated and effec-tive at transporting charge have a shallow conductance decay and, consequently, a low value. Here, we limit the discussion to values obtained from measurements of single molecules (rather than molecular assemblies), in which conductance is dominated by coherent tunnelling mechanisms. Excellent reviews have been written that discuss values obtained from a wider range of meas-urement techniques, wire structures and conductance mechanisms96,97.

Representative values for several oligomeric mate-rials that conduct via coherent tunnelling are listed in TABLE1. Alkanes are characterized by high values (0.84 1) because they do not have strongly conjugated bonds that can carry charge12,57. Permethyloligosilanes ([SiMe2]n) terminated with methylthiomethyl electrode linkers have a value (0.39 1) comparable to that of aromatic conductors48. Despite their structural sim-ilarity to alkanes, oligosilanes transport charge more effectively because SiSi bonds are more strongly con-jugated than CC bonds, as their bonding orbitals are much larger in size and much higher in energy98. From the perspective of the transmission function (BOX1), the nearest neighbour coupling in silanes is stronger than in alkanes. The low value in silanes also opens up the possibility of observing QI effects in systems, as discussed further below. Isostructural permethyloligo-germanes also have a low value (0.36 1), which is slightly lower than that of permethyloligosilanes49.

-conjugated backbones tend to be associated with low values. Conjugated but non-aromatic systems typically have a lower value than purely aromatic species. Mapping the conductance of molecular wires against their degree of aromaticity suggests that the conductance is inversely proportional to the resonance stabilization energy99. In the transmission picture, this aromatic stabilization energy decreases the on-site energies (), thereby lowering E (BOX1). The higher

R E V I E W S



conductance of non-aromatic structures is supported by bulk-scale conductance100 and optoelectronic101 experiments, which show that polymer backbones with strong quinoidal character (and consequently weak aromatic character) are superior at delocalizing charge along the backbone. Anti-aromaticity describes the energetic destabilization that arises from the inter-action of (4n) electrons in a planar ring. These destabilized structures should possess a high HOMO energy, and it has been predicted that anti-aromatic compounds will exhibit high conductance102. However, the instability of purely anti-aromatic structures has precluded their study inSMEs.

Substituent effects on conductance. Synthetic mod-ification of the substituent structure is another han-dle for altering the charge-transport properties of the molecular wire. The relationship between substituent electronics and conductance was explored in 1,4-diamin-obenzene-based molecular wires, in which different sub-stituents were introduced at the 2-, 3-, 5- and 6- positions. It was determined that, in this particular system, elec-tron-donating groups (such as Me and OMe) increase conductance, whereas electron-withdrawing groups

(such as CF3, Br, Cl and F) decrease conductance103. These trends arise from the HOMO-conducting nature of the amine linkers; substituents that push electron density into the benzene ring raise the HOMO energy towards the EF of gold. Electron-withdrawing substituents sta-bilize the backbone and lower the HOMO energy away from the EF. These findings can be analysed in the context of physical organic chemistry by constructing a Hammett plot from molecular conductance and the Hammett constants (para)104 (FIG.3d). Hammett constants were originally quantified to correlate the electron-withdraw-ing (para) or electron-donating (+para) nature of paralinked aromatic substituents with chemical reactivity. This relationship between para and conductance, demon-strated in FIG.3d, suggests that, just as chemical reaction rates depend on the differences in energy between the ground state and the transition state, conductance is related to the energy difference between the metalmol-ecule hybridized conducting orbital and the EF. The same general principle is observed in single-molecule thermoelectric devices based on benzene dithiol junc-tions, in which electron-donating substituents increase the Seebeck coefficient (hole-transporting character) and electron-withdrawing substituents decreaseit105.

Table 1 | values of oligomeric materials with conductance dominated by coherent-tunnelling mechanisms

Backbone Structure (1) Refs

Alkane

n0.84 57

Silane

Si

n

0.39 48

Germane

Ge

n

0.36 49

Alkene

n

0.22 145

Alkynen

0.170.32 44

p-phenylene

n

0.43 13

p-phenylene ethynylene

n

0.20, 0.34 146,147

Thiophene S

n

0.16, 0.4 138,148

Thiophene dioxideS

O O

n

0.20 41

Cyclopenta-difluorene

n

0.21 149

R E V I E W S



Various research groups have studied the confor-mationconductance relationship in biphenyl rings by synthetically modifying the 2,2 substituents to alter the torsion angle between the two phenyl rings13,106,107. Regardless of whether the biphenyl rings are termi-nated by - or *-conducting anchors, conductance decreases as the torsion angle () increases and the

conjugation between the rings weakens (FIG.3e). In ref-erence to the transmission picture described in BOX1, the coupling between neighbouring sites decreases with cos(). These examples illustrate how synthetic variations of the substituent group can alter conduct-ance by tuning the electronics or stereochemistry of the molecular backbone.


a

b

c

d

100

102

108

106

104

1010

0.0

1.0

2.0

3.0

1.6

0.8

0.0

0.8

1.6

2

2

1

1

0

0

1

1

2

2

0 25 50 75 100

0 200 200 400 600 800 400

Time (s)

Potential (mV)

Energy EF (eV)

Ener

gy

EF (

eV)

104

103

0.01

Erase pulse: Verase

= 1.6 V

Write pulse: Vwrite

= +1.6 V Read pulse: Vread

= +1.1 V

NH2

H2NH2N NH2

pH = 3pH = 8

Volt

age

(V)

Con

duct

ance

(G0)

Cur

rent

(nA

)Tr

ansm

issi

on

AC-DT

AQ-DT

AQ-MT

On

Off

Reading time 3 s

Couplingenergy

+2.4 eV +0.6 eV

N1

N2

4.6 eV

N1

+ N2

7.0 eV 5.6 eV

5.0 eV

Lone pair coupling strength

Figure 4 | Single-molecule conductance switches and quantum interference features. a|Reversible voltage-induced switching between on and off states in an oligo(phenylene ethynylene)-based wire. b|Calculated transmission for monothiolated anthraquinone (AQ-MT), dithiolated anthraquinone (AQ-DT) and dithiolated anthracene (AC-DT). The diagram on the right depicts the energy levels (from top to bottom) of the lowest unnoccupied molecular orbital (LUMO), the highest occupied molecular orbital (HOMO) and the HOMO-1 of AC-DT, AQ-DT and AQ-MT. AQ-MT and AQ-DT exhibit anti-resonance features. c| The plotted points represent conductance peak values at particular gate potentials and pH values for an anthracene and anthraquinone-based quantum interference switch. d|Isosurfaces of the two highest occupied molecular orbitals containing lone pair character that exhibit anti-symmetric (N1 N2) and symmetric (N1 + N2) lone pair phases (B3LYP/6-31G**). The difference in energy between these two states corresponds to the coupling strength between the two lone pairs. Molecules with strongly coupled lone pairs typically do not demonstrate anti-resonance quantum interference features near the Fermi level, EF. Panel a is reproduced with permission from REF.118, WileyVCH. Panel b is from REF.123, Nature Publishing Group. Panel c is adapted with permission from REF.125, WileyVCH.

R E V I E W S



Emerging areas in single-molecule electronicsIn the previous section, we explored how altering the molecular wire structure and stereochemistry can affect the fundamental charge-transport properties of the single-molecule junction. In this section, we discuss how the structural units of the junction (par-ticularly the anchor and the bridge) can be manipu-lated to create new SME devices. Many sophisticated switching platforms made of ensembles of molecules have been developed and are thoroughly discussed in other reviews5,96,108,109; here, we focus specifically on the opportunities and challenges in the development of single-molecule devices that switch conductance reversibly and on demand between stable onoff states. We then discuss how QI, an emerging phenomenon specific to molecular-scale devices, depends on the strength of electronic communication between the anchor groups through the bridge. We conclude by rationalizing QI from a chemical perspective.

Incorporating switching functions into the junction. There are many examples of multimodal switching from solution-phase chemistry, in which the molecule can switch between different isomers using optical, electrochemical or pH triggers. However, the ability of a molecule to switch in solution is not always retained when it is bound to electrodes. First, the activation barrier for switching needs to be high enough so that the switching can be induced predictably and will not occur randomly. For example, it was demonstrated in a bianthrone-based device that even if the barrier between bistable states is high for the free molecule in solution, the barrier between these states can lower as the molecular orbitals hybridize with the orbitals of the electrode and cause switching to occur uncontrollably even at low temperatures110.

Second, holding the molecule at a fixed distance between two electrodes can prevent switching in sys-tems in which the two states have different anchor- to-anchor lengths. For this reason, there are many studies on switching between cis and trans isomers in stilbene and azobenzene ensembles in which the mole-cules only need to be bound at one electrode; however, controlled switching between cis and trans isomers in a single metalmoleculemetal junction has yet to be reported.

Last, and more specific to photoswitching, if the coupling between the electrode and a photoactive molecule is strong, the photoexcited state of the mol-ecule can transfer charge to the electrode rather than photoisomerizing. This mechanism is thought to be the reason why certain diarylethene switches, despite hav-ing on- and off-state geometries of similar molecular length, do not operate reversibly in break-junction set-ups111. Such problems can potentially be circumvented by synthesizing wire structures in which the backbone is decoupled from the electrode112.

It is challenging to create a system in which switch-ing occurs in a single metalmoleculemetal junc-tion. Although many examples of switching triggered by pH113,114 and light112,115 have been reported, they

typically do not describe switching within the same individual molecule. Mechanically triggered switch-ing22,25,48, in which the stretching or compressing of the junction induces the switching, and voltage-triggered switching116118 (FIG.4a) have been used to successfully switch conductance in individual junctions. However, this approach is difficult to implement in devices in which the electrodes must be stationary or the volt-age bias between two terminals must be held constant. There is much potential for growth in this branch of SMEs; innovations in synthesis and engineering will promote the development of single-junction conduct-ance switches, which are crucial for the realization of single-molecule logic and memory components.

Quantum interference effects. The concept of QI was first developed to describe the wave interference of photons and electrons in the context of double-slit experiments. In the SMEs community, the term quan-tum interference has been appropriated to describe the interference between the electron waves propagat-ing through molecular orbitals in a single-molecule junction119121. QI-based devices have captured the imagination of many in the SMEs community because there are no parallels in conventional solid-state elec-tronics. Constructive interference enhances conduct-ance, whereas destructive interference suppresses conductance. Constructive QI has been reported in a double-backbone junction, in which the observed conductance was found to exceed the conductance of two single-backbone junctions operated in parallel122. The SMEs field has focused its attention mostly on destructive QI, which is characterized by a sharp dip or an anti-resonance in the transmission probability of non-equilibrium Greens function calculations (FIG.4b) or in differential conductance experiments123,124. These anti-resonances are most relevant for single-molecule conductance when they occur near the electrode EF (but can occur anywhere in the energy landscape).

When destructive interference occurs near the EF of the electrode, little or no conductance is observed. This is appealing because the destructive QI path-way, if it can be switched on demand, might serve as an off state that is exceedingly low in conductance; thus, QI offers a potential route for the creation of molecular devices with large on/off ratios. Many the-oretical devices119,120 of this kind have been imagined, in which the device can either gate the position of the anti-resonance or switch the transport pathway through the molecule between non-QI and destruc-tive-QI pathways however, few of these devices have been experimentally realized. Recently, several groups have created QI switches in which electrochemical triggers induce bulk switching between cross-con-jugated anthra quinone (QI) and linearly conjugated dihydroxy anthracene (non-QI) backbone structures that differ in conductance by more than an order of magnitude125,126 (FIG.4c).

The notion of destructive QI is quite familiar to chemists, although from a different perspective. For example, para-substituted phenyl rings yield high

R E V I E W S




SLL

AuAu

LLAuAu

Me Me

a Atom-counting approach b Resonance-structure analysis

c Non-equivalent pathways

e Influence of aromaticity

d Examples of resonance-structure analysis

f Odd non-alternant hydrocarbons

SLL

Au Au

LL

MeMe

Au Au

NNH

N HN

LLAu Au

LLAuAu

L

L Au

AuQI

No QI

no QI

NNH

N HN

LL

Au

Au

NNH

N HN

LL

Au

Au

L

L Au

Au

O

O

L

L

Au

Au

QI

O

O

L

L

Au

Au

OH

OH

L

LAu

Au

No QI

OH

OH

L

L

Au

Au

HNN

NH N

L

LAu

Au

QI

HNN

NH N

L

L

Au

Au

L

L

Au

Au

No QIL

L Au

Au

QI

LL

Au

Au

4.8 x 104 G0

1.8 x 104 G0

L

L

Au

Au

Anti-aromatic

32 x 105 G0

1,3-azulene

Au

Au Au

L

L

Au

2 x 105 G0

5,7-azulene

LL L

L

PyrenePorphyrin

DihydroxyanthraceneAnthraquinone

L L

32 x 105 G0

2,6-azulene

Figure 5 | Pen-and-paper methods to predict quantum interference. a|In the atom-counting approach, a continuous path is drawn from one linker to the other. A circle is drawn around neighbouring-atom pairs that are off path, as in the para case. If there are single atoms that are off path and do not have an off-path neighbour, as in the meta case, destructive quantum interference (QI) is expected to occur. b|Resonance-structure analysis to predict QI. c|When there are two non-equivalent pathways, as in the pyrrole ring, some arrow-pushing pathways are dead ends (left panel). A good pathway is also shown (right panel). d|Resonance-structure analysis of QI in porphyrin, pyrene, anthraquinone and dihydroxyanthracene. e|Conductance peak values and resonance structures for 2,5-disubstituted cyclopentadiene (top panel) and thiophene (bottom panel). The conductive quinoidal resonance form makes a more important contribution to the electronic structure of the cyclopentadiene compared with that of the thiophene because the former is non-aromatic. f|Conductance peak values and resonance structures of azulene constitutional isomers. Both 1,3-azulene and 5,7-azulene isomers show strong anti-resonances in their calculated transmission functions, but the anti-resonance is shifted towards a higher energy for 1,3-azulene its transmission is then much higher at the Fermi level, EF. The 1,3-azulene isomer demonstrates an anti-aromatic resonance structure, whereas the 5,7-azulene isomer does not. The 2,6-isomer is as conductive as the 1,3-isomer, even though it does not have a QI feature near the EF.

R E V I E W S



conductance values, whereas metasubstituted phenyl rings demonstrate substantially attenuated conductance owing to destructive interference pathways40,127. This destructive QI ultimately arises from the node at the meta position in the HOMO that prevents electronic coupling between two meta substituents. Therefore, metalinked wires show poor conductance and destruc-tive QI for the same reason that electrophilic aromatic substitution reactions with electron-rich substituents, in general, do not occur at the meta-position. Cross-conjugated structures, such as the anthraquinone mol-ecule in the example above, often show anti-resonances in their transmission probability calculations128 for essentially the same reason the two primary unsatu-rated pathways are not conjugated with each other. The lack of communication between the two unsaturated ends of a cross-conjugated molecule has been well established in the context of chemical reactivity and spectroscopy129.

The coupling strength between the two anchor groups through the bridge is thus strongly related to QI and conductance. This coupling strength can be qualitatively evaluated from ground-state density functional theory computations that model the elec-tronic structure of the molecule, as shown in FIG.4d for 1,3- and 1,4-diaminobenzene. Our analysis is limited to the two highest occupied molecular orbitals that contain the lone pair orbitals. The phases of the two lone pairs (N1 and N2) are typically symmetric in one of these molecular orbitals (N1 + N2) and anti-sym-metric in the other (N1 N2). The energy needed to decouple the lone pairs can be estimated by consid-ering the energy cost of adding or subtracting these molecular orbitals from each other, which would localize the orbital density on N1 or N2, respectively. The energy difference between these two molecular orbitals is 2.4 eV for 1,4-diaminobenzene and 0.6 eV for 1,3-diaminobenzene; therefore, coupling is signifi-cantly stronger in the paralinked molecule than in the metalinked molecule. This analysis can be extended to other simple aromatic systems to rationalize con-ductivity trends that arise from varying linker connec-tivity or backbone composition130.

Several other methods have been developed to pre-dict QI in alternant -conjugated systems. With orbital symmetry rules, it is possible to predict whether QI occurs in naphthalene systems with different linker connectivity points131. The orbital-symmetry approach is based on a qualitative comparison of the phases of the ring carbon orbitals in the HOMO and in the LUMO. This methodology was also used to elucidate the rela-tionship between switching ratio and anchor-group connectivity in photoactive molecules132.

Pen-and-paper methods for predicting quantum interference. Simple graphical methods, such as the star labelling formalisms of orbital pairing in molec-ular orbital theory133, can be used to predict whether even-alternant hydrocarbons demonstrate alike (QI) or disjoint (non-QI) coupling128. Atom-counting approaches can also be used to forecast destructive QI

in cyclic and acyclic conjugated molecules134 (FIG.5a). These graphical approaches are useful because only a pen and paper is required to quickly predict whether there is QI in a molecular structure. These methods should be familiar to reaction chemists because they are essentially shorthand for arrow pushing in -conjugated resonance forms. It is then reasonable that QI can be predicted by simply drawing and analysing resonance structures.

Resonance structures are simple representations of the -electron interactions in conjugated molecules. No single resonance structure is entirely descriptive of the electronic character of a -conjugated molecule, and the molecule does not interconvert between reso-nance structures rather, the electronic structure is a weighted average of resonance forms. The rules and conventions for drawing reasonable resonance struc-tures by arrow pushing can be found in any introduc-tory organic chemistry textbook. Resonance logic can be used to predict QI in -conjugated molecules (FIG.5b) by the following steps. First, draw the chem-ical structure of the molecular bridge with Au atoms attached to each linker (L) atom. Second, attempt to draw a reasonable resonance structure that delocal-izes one LAu bond onto the other Au atom by using arrow-pushing formalisms. If it is possible to draw a resonance structure where there is a quinoidal structure in the molecular bridge, a positive charge on one Au atom and a negative charge on the other Au atom, QI should not occur. If there are non-equivalent pathways in an aromatic ring, the choice of pathway is important, because some pathways do not connect the Au atoms (FIG.5c). All pathways must be sampled before determin-ing whether there is QI in the structure. The reasonable range of resonance structures used in this approach is limited to those that possess a single positive and a sin-gle negative charge. This ensures that only the fate of the original LAu bond is tracked. Last, if there are no resonance structures that can be drawn that meet the criteria in the second step, the molecule should demon-strate destructiveQI.

According to these rules, QI essentially depends on whether it is possible to draw a resonance structure of the junction in which one linkerelectrode bond delo-calizes through the -conjugated bridge onto the other electrode atom. This is reasonable because conductance describes how well one electrode is connected with the other through the conjugated molecule. The charges that are drawn on the Au atoms in non-QI resonance structures are symbolic of charge transfer between the electrodes. If there is no resonance form that connects the Au electrodes in a single path through the sys-tem, destructive QI occurs. FIGURE5d shows that this approach holds for various systems such as porphy-rins135,136, pyrenes137, anthraquinones123 and anthra-cenes123. For practice, readers can try this method on various other polycyclic aromatic hydrocarbons and check the answers against the corresponding trans-mission calculations138.

The resonance-structure approach is valuable because it is compatible with chemical intuition. It can

R E V I E W S



be used not only to predict QI but also to explain why aromaticity tends to diminish charge-transport ability. FIGURE5e depicts the aromatic and quinoidal resonance structures of 2,5-disubstituted cyclopentadiene and thiophene99. The quinoidal resonance form optimizes coupling between the electrodes, as stipulated by the resonance-structure analysis above. However, aromatic wires have weak quinoidal character because of the strong stabilizing energy associated with aromaticity. Conductance is thus higher for nonaromatic wires because their electronic structures are more quinoidal in character.

In odd non-alternant hydrocarbons, such as azu-lene, QI pathways can still result in strong conduct-ance46. The 1,3- and 5,7-disubstituted azulene isomers both demonstrate transmission anti-resonances, but the 1,3-isomer gives a much higher conductance (FIG.5f) owing to the shift of its anti-resonance position towards a higher energy. The 1,3-isomer is as conduc-tive as the 2,6-isomer that does not have a QI feature near the EF. If we draw the reasonable resonance struc-tures for the 1,3- and 5,7-isomers, it becomes imme-diately clear that the 1,3-isomer has an anti-aromatic resonance form, whereas the 5,7-isomer does not. Recalling that anti-aromaticity is expected to result in enhanced conductance, it can be understood why the 1,3-isomer has the higher conductance despite the poor coupling between the anchor groups through the conductive pathway.

Despite the speed and the ease of these pen-and- paper approaches, there are also limitations. They cannot be used to quantitatively predict the conduct-ance value or the position of the anti-resonance peak in transmission calculations. Furthermore, they can only be used to predict QI in -conducting systems; simple methods for predicting QI in -conjugated sys-tems, such as silanes, do not yet exist. More sophis-ticated methods must be used in these instances139. However, these simple methods are powerful because they enable rapid prediction of whether a molecular structure will exhibit QI. They can be used to quickly draw the resonance structures of many complex -conjugated structures and to rapidly screen which of these compounds might possess interesting or anomalous QI properties.

PerspectivesIn all realms of science, great progress can be made when lessons from the past are applied to understand the challenges and opportunities of the present and future. The ideas at the base of the emerging field of SMEs are rooted in the same principles developed in chemistry over the past century. In this Review, we have discussed the benefits and opportunities that arise from using chemical principles to guide the design and understanding of functional molecular electronic components.

Concepts borrowed from reaction chemistry con-tinue to inspire new SME devices but, at the same time, discoveries in SMEs produce new knowledge in chemistry, engineering and other fields of science. Research in SMEs, particularly through scanning probe measurements on self-assembled monolayers, has shed light on how molecules assemble and behave when adsorbed or bonded to metal surfaces53,140. Extrapolating molecular conductance from these densely packed monolayer assemblies provides infor-mation on how molecules communicate with one another and with their surroundings in thin molecu-lar films52. Single-molecule QI concepts have recently been used to design new ligand dyes for dye-sensitized solar cells and new dielectric materials141,142. The dis-coveries made in the context of SMEs are also relevant for current and future integrated circuit technologies. The insights gained from studying conductance at the molecular scale will help us to understand how the charge-transport properties of semiconducting materi-als might change as their dimensions decrease and they begin to resemble molecules rather than bulksolids.

The past two decades have seen a vast deepening in the knowledge of how to create, measure and under-stand single-molecule junctions. One of the immediate challenges is how to use this newly gained knowledge to incorporate these structures into system-level devices. These new nanoscale architectures and inter-connections could revolutionize nanoscale electronics by providing access to the vast store of structures avail-able through synthetic chemistry. The combination of synthetic chemistry and electronics is an inspiration for the creation of fields and applications yet to be imagined.

1. Aviram,A. & Ratner,M.A. Molecular rectifiers. Chem. Phys. Lett. 29, 277283 (1974).

2. Choi,H. & Mody,C.C.M. The long history of molecular electronics: microelectronics origins of nanotechnology. Soc. Stud. Sci. 39, 1150 (2009).

3. Tao,N.J. Electron transport in molecular junctions. Nat. Nanotechnol. 1, 173181 (2006).

4. Moth-Poulsen,K. & Bjrnholm,T. Molecular electronics with single molecules in solid-state devices. Nat. Nanotechnol. 4, 551556 (2009).

5. Aradhya,S.V. & Venkataraman,L. Single-molecule junctions beyond electronic transport. Nat. Nanotechnol. 8, 399410 (2013).

6. Ward,D.R., Scott,G.D., Keane,Z.K., Halas,N.J. & Natelson,D. Electronic and optical properties of electromigrated molecular junctions. J.Phys. Condens. Matter 20, 374118 (2008).

7. Bumm,L.A. etal. Are single molecular wires conducting? Science 271, 17051707 (1996).

8. Moreland,J. & Ekin,J.W. Electron tunneling experiments using NbSn break junctions. J.Appl. Phys. 58, 38883895 (1985).

9. Muller,C.J., van Ruitenbeek,J.M. & de Jongh,L.J. Experimental observation of the transition from weak link to tunnel junction. Phys. C Supercond. 191, 485504 (1992).

10. Smit,R.H.M. etal. Measurement of the conductance of a hydrogen molecule. Nature 419, 906909 (2002).

11. Reed,M.A. Conductance of a molecular junction. Science 278, 252254 (1997).

12. Xu,B. & Tao,N.J. Measurement of single-molecule resistance by repeated formation of molecular junctions. Science 301, 12211223 (2003).

13. Venkataraman,L., Klare,J.E., Nuckolls,C., Hybertsen,M.S. & Steigerwald,M.L. Dependence of single-molecule junction conductance on molecular conformation. Nature 442, 904907 (2006).

14. Wold,D.J. & Frisbie,C.D. Formation of metalmoleculemetal tunnel junctions: microcontacts to alkanethiol monolayers with a conducting AFM tip. J.Am. Chem. Soc. 122, 29702971 (2000).

15. Park,H., Lim,A.K.L., Alivisatos,A.P., Park,J. & McEuen,P.L. Fabrication of metallic electrodes with nanometer separation by electromigration. Appl. Phys. Lett. 75, 301303 (1999).

16. Andres,R.P. etal. Self-assembly of a two-dimensional superlattice of molecularly linked metal clusters. Science 273, 16901693 (1996).

17. Cui,X.D. etal. Reproducible measurement of single-molecule conductivity. Science 294, 571574 (2001).

18. Guo,X. etal. Covalently bridging gaps in single-walled carbon nanotubes with conducting molecules. Science 311, 356359 (2006).

19. Prins,F. etal. Room-temperature gating of molecular junctions using few-layer graphene nanogap electrodes. Nano Lett. 11, 46074611 (2011).

R E V I E W S



20. Morpurgo,A.F., Marcus,C.M. & Robinson,D.B. Controlled fabrication of metallic electrodes with atomic separation. Appl. Phys. Lett. 74, 20842086 (1999).

21. Chen,X. etal. Chemical fabrication of heterometallic nanogaps for molecular transport junctions. Nano Lett. 9, 39743979 (2009).

22. Quek,S.Y. etal. Mechanically controlled binary conductance switching of a single-molecule junction. Nat. Nanotechnol. 4, 230234 (2009).

23. Su,T.A. etal. Silicon ring strain creates high-conductance pathways in single-molecule circuits. J.Am. Chem. Soc. 135, 1833118334 (2013).

24. Kiguchi,M. etal. Single molecular resistive switch obtained via sliding multiple anchoring points and varying effective wire length. J.Am. Chem. Soc. 136, 73277332 (2014).

25. Moreno-Garca,P. etal. Charge transport in C60-based dumbbell-type molecules: mechanically induced switching between two distinct conductance states. J.Am. Chem. Soc. 137, 23182327 (2015).

26. Lafferentz,L. etal. Conductance of a single conjugated polymer as a continuous function of its length. Science 323, 11931197 (2009).

27. Meisner,J.S. etal. A single-molecule potentiometer. Nano Lett. 11, 15751579 (2011).

28. Leary,E. etal. Unambiguous one-molecule conductance measurements under ambient conditions. Nano Lett. 11, 22362241 (2011).

29. Martin,C.A. etal. Fullerene-based anchoring groups for molecular electronics. J.Am. Chem. Soc. 130, 1319813199 (2008).

30. Aradhya,S.V., Frei,M., Hybertsen,M.S. & Venkataraman,L. Van der Waals interactions at metal/organic interfaces at the single-molecule level. Nat. Mater. 11, 872876 (2012).

31. Heimel,G. etal. Charged and metallic molecular monolayers through surface-induced aromatic stabilization. Nat. Chem. 5, 187194 (2013).

32. Schlfer,H.L. & Gliemann,G. Basic Principles of Ligand Field Theory (Wiley-Interscience, 1969).

33. Venkataraman,L. etal. Single-molecule circuits with well-defined molecular conductance. Nano Lett. 6, 458462 (2006).

34. Reddy,P., Jang,S.-Y., Segalman,R.A. & Majumdar,A. Thermoelectricity in molecular junctions. Science 315, 15681571 (2007).

35. Widawsky,J.R., Darancet,P., Neaton,J.B. & Venkataraman,L. Simultaneous determination of conductance and thermopower of single molecule junctions. Nano Lett. 12, 354358 (2012).

36. Derosa,P.A. & Seminario,J.M. Electron transport through single molecules: scattering treatment using density functional and green function theories. J.Phys. Chem. B 105, 471481 (2001).

37. Kamenetska,M. etal. Conductance and geometry of pyridine-linked single-molecule junctions. J.Am. Chem. Soc. 132, 68176821 (2010).

38. Lee,W. etal. Heat dissipation in atomic-scale junctions. Nature 498, 209212 (2013).

39. Koga,J., Tsuji,Y. & Yoshizawa,K. Orbital control of single-molecule conductance perturbed by -accepting anchor groups: cyanide and isocyanide. J.Phys. Chem. C 116, 2060720616 (2012).

40. Mayor,M. etal. Electric current through a molecular rod relevance of the position of the anchor groups. Angew. Chem. Int. Ed. Engl. 42, 58345838 (2003).

41. Dell,E.J., Capozzi,B., Xia,J., Venkataraman,L. & Campos,L.M. Molecular length dictates the nature of charge carriers in single-molecule junctions of oxidized oligothiophenes. Nat. Chem. 7, 209214 (2015).

42. Li,Z., Smeu,M., Ratner,M.A. & Borguet,E. Effect of anchoring groups on single molecule charge transport through porphyrins. J.Phys. Chem. C 117, 1489014898 (2013).

43. Park,Y.S. etal. Frustrated rotations in single-molecule junctions. J.Am. Chem. Soc. 131, 1082010821 (2009).

44. Moreno-Garca,P. etal. Single-molecule conductance of functionalized oligoynes: length dependence and junction evolution. J.Am. Chem. Soc. 135, 1222812240 (2013).

45. Batra,A. etal. Tuning rectification in single-molecular diodes. Nano Lett. 13, 62336237 (2013).

46. Xia,J. etal. Breakdown of interference rules in azulene, a nonalternant hydrocarbon. Nano Lett. 14, 29412945 (2014).

47. Danilov,A. etal. Electronic transport in single molecule junctions: control of the moleculeelectrode coupling through intramolecular tunneling barriers. Nano Lett. 8, 15 (2008).

48. Su,T.A., Li,H., Steigerwald,M.L., Venkataraman,L. & Nuckolls,C. Stereoelectronic switching in single-molecule junctions. Nat. Chem. 7, 215220 (2015).

49. Su,T.A. etal. Single-molecule conductance in atomically precise germanium wires. J.Am. Chem. Soc. 137, 1240012405 (2015).

50. Arroyo,C.R. etal. Influence of binding groups on molecular junction formation. J.Am. Chem. Soc. 133, 1431314319 (2011).

51. Li,H. etal. Electric field breakdown in single molecule junctions. J.Am. Chem. Soc. 137, 50285033 (2015).

52. Tour,J.M. etal. Self-assembled monolayers and multilayers of conjugated thiols, ,-dithiols, and thioacetyl-containing adsorbates. Understanding attachments between potential molecular wires and gold surfaces. J.Am. Chem. Soc. 117, 95299534 (1995).

53. Hkkinen,H. The goldsulfur interface at the nanoscale. Nat. Chem. 4, 443455 (2012).

54. Gonzlez,M.T. etal. Break-junction experiments on acetyl-protected conjugated dithiols under different environmental conditions. J.Phys. Chem. C 115, 1797317978 (2011).

55. Haiss,W. etal. Precision control of single-molecule electrical junctions. Nat. Mater. 5, 9951002 (2006).

56. Haiss,W. etal. Impact of junction formation method and surface roughness on single molecule conductance. J.Phys. Chem. C 113, 58235833 (2009).

57. Li,X. etal. Conductance of single alkanedithiols: conduction mechanism and effect of moleculeelectrode contacts. J.Am. Chem. Soc. 128, 21352141 (2006).

58. Chen,F., Li,X., Hihath,J., Huang,Z. & Tao,N. Effect of anchoring groups on single-molecule conductance: comparative study of thiol-, amine-, and carboxylic-acid-terminated molecules. J.Am. Chem. Soc. 128, 1587415881 (2006).

59. Li,C. etal. Charge transport in single Au/alkanedithiol/Au junctions: coordination geometries and conformational degrees of freedom. J.Am. Chem. Soc. 130, 318326 (2008).

60. Inkpen,M.S. etal. New insights into single-molecule junctions using a robust, unsupervised approach to data collection and analysis. J.Am. Chem. Soc. 137, 99719981 (2015).

61. Rascn-Ramos,H., Arts,J.M., Li,Y. & Hihath,J. Binding configurations and intramolecular strain in single-molecule devices. Nat. Mater. 14, 517522 (2015).

62. Li,Z. etal. Hapticity-dependent charge transport through carbodithioate-terminated [5,15-bis(phenylethynyl)porphinato]zinc(II) complexes in metalmoleculemetal junctions. Nano Lett. 14, 54935499 (2014).

63. Batra,A. etal. Trimethyltin-mediated covalent goldcarbon bond formation. J.Am. Chem. Soc. 136, 1255612559 (2014).

64. Kim,Y., Pietsch,T., Erbe,A., Belzig,W. & Scheer,E. Benzenedithiol: a broad-range single-channel molecular conductor. Nano Lett. 11, 37343738 (2011).

65. Ulrich,J. etal. Variability of conductance in molecular junctions. J.Phys. Chem. B 110, 24622466 (2006).

66. Xiao,X., Xu,B. & Tao,J. Measurement of single molecule conductance: benzenedithiol and benzenedimethanedithiol. Nano Lett. 4, 267271 (2004).

67. Cheng,Z.-L. etal. In situ formation of highly conducting covalent AuC contacts for single-molecule junctions. Nat. Nanotechnol. 6, 353357 (2011).

68. Kaletov,E. etal. The scope of direct alkylation of gold surface with solutions of C1C4 n-alkylstannanes. J.Am. Chem. Soc. 137, 1208612099 (2015).

69. Khobragade,D. etal. Preparation of covalent long-chain trialkylstannyl and trialkylsilyl salts and an examination of their adsorption on gold. Langmuir 26, 84838490 (2010).

70. Chen,W. etal. Highly conducting -conjugated molecular junctions covalently bonded to gold electrodes. J.Am. Chem. Soc. 133, 1716017163 (2011).

71. Schwarz,F. etal. High-conductive organometallic molecular wires with delocalized electron systems strongly coupled to metal electrodes. Nano Lett. 14, 59325940 (2014).

72. Hong,W. etal. Trimethylsilyl-terminated oligo(phenylene ethynylene)s: an approach to single-molecule junctions with covalent AuC -bonds. J.Am. Chem. Soc. 134, 1942519431 (2012).

73. Wuts,P.G.M. & Greene,T.W. Greenes Protective Groups in Organic Synthesis 4th edn (Wiley, 2006).

74. Nagahara,L.A., Thundat,T. & Lindsay,S.M. Preparation and characterization of STM tips for electrochemical studies. Rev. Sci. Instrum. 60, 31283130 (1989).

75. Saby,C., Ortiz,B., Champagne,G.Y. & Be,D. Electrochemical modification of glassy carbon electrode using aromatic diazonium salts. 1. Blocking effect of 4-nitrophenyl and 4-carboxyphenyl groups. Langmuir 7463, 68056813 (1997).

76. Hines,T. etal. Controlling formation of single-molecule junctions by electrochemical reduction of diazonium terminal groups. J.Am. Chem. Soc. 135, 33193322 (2013).

77. Patai,S. Diazonium Diazo Groups Vol. 1 (Wiley, 1978).78. Pearson,R.G. Hard and soft acids and bases. J.Am.

Chem. Soc. 85, 35333539 (1963).79. Ullmann,F. & Bielecki,J. Ueber synthesen in der

biphenylreihe. Ber. Dtsch. Chem. Ges. 34, 21742185 (1901).

80. Cai,J. etal. Atomically precise bottom-up fabrication of graphene nanoribbons. Nature 466, 470473 (2010).

81. Koch,M., Ample,F., Joachim,C. & Grill,L. Voltage-dependent conductance of a single graphene nanoribbon. Nat. Nanotechnol. 7, 713718 (2012).

82. Nguyen,S. & Grubbs,R.H. Syntheses and activities of new single-component, ruthenium-based olefin metathesis catalysts. J.Am. Chem. Soc. 115, 98589859 (1993).

83. Tulevski,G.S., Myers,M.B., Hybertsen,M.S., Steigerwald,M.L. & Nuckolls,C. Formation of catalytic metalmolecule contacts. Science 309, 591594 (2005).

84. Kim,T., Vzquez,H., Hybertsen,M.S. & Venkataraman,L. Conductance of molecular junctions formed with silver electrodes. Nano Lett. 13, 33583364 (2013).

85. Kaneko,S., Nakazumi,T. & Kiguchi,M. Fabrication of a well-defined single benzene molecule junction using Ag electrodes. J.Phys. Chem. Lett. 1, 35203523 (2010).

86. Haick,H., Ghabboun,J. & Cahen,D. Pd versus Au as evaporated metal contacts to molecules. Appl. Phys. Lett. 86, 20042006 (2005).

87. Kiguchi,M. etal. Highly conductive molecular junctions based on direct binding of benzene to platinum electrodes. Phys. Rev. Lett. 101, 046801 (2008).

88. Kim,Y. etal. Conductance and vibrational states of single-molecule junctions controlled by mechanical stretching and material variation. Phys. Rev. Lett. 106, 196804 (2011).

89. Yelin,T. etal. Atomically wired molecular junctions: connecting a single organic molecule by chains of metal atoms. Nano Lett. 13, 19561961 (2013).

90. Ko,C.H., Huang,M.J., Fu,M.D. & Chen,C.H. Superior contact for single-molecule conductance: electronic coupling of thiolate and isothiocyanate on Pt, Pd, and Au. J.Am. Chem. Soc. 132, 756764 (2010).

91. Seminario,J.M., De La Cruz,C.E. & Derosa,P.A.A. Theoretical analysis of metalmolecule contacts. J.Am. Chem. Soc. 123, 56165617 (2001).

92. Adak,O., Korytr,R., Joe,A.Y., Evers,F. & Venkataraman,L. Impact of electrode density of states on transport through pyridine-linked single molecule junctions. Nano Lett. 15, 37163722 (2015).

93. Schmaus,S. etal. Giant magnetoresistance through a single molecule. Nat. Nanotechnol. 6, 185189 (2011).

94. Wende,H. etal. Substrate-induced magnetic ordering and switching of iron porphyrin molecules. Nat. Mater. 6, 516520 (2007).

95. Iacovita,C. etal. Visualizing the spin of individual cobaltphthalocyanine molecules. Phys. Rev. Lett. 101, 4043 (2008).

96. Metzger,R.M. Unimolecular Electronics. Chem. Rev. 115, 50565115 (2015).

97. Luo,L., Choi,S.H. & Frisbie,C.D. Probing hopping conduction in conjugated molecular wires connected to metal electrodes. Chem. Mater. 23, 631645 (2011).

98. Miller,R.D. & Michl,J. Polysilane high polymers. Chem. Rev. 89, 13591410 (1989).

99. Chen,W. etal. Aromaticity decreases single-molecule junction conductance. J.Am. Chem. Soc. 136, 918920 (2014).

100. Wudl,F., Kobayashi,M. & Heeger,A.J. Poly(isothianaphthene). J.Org. Chem. 49, 33823384 (1984).

101. Douglas,J.D. etal. Functionalized isothianaphthene monomers that promote quinoidal character in donoracceptor copolymers for organic photovoltaics. Macromolecules 45, 40694074 (2012).

R E V I E W S



102. Breslow,R. & FossJr,F.W. Charge transport in nanoscale aromatic and antiaromatic systems. J.Phys. Condens. Matter 20, 374104 (2008).

103. Venkataraman,L. etal. Electronics and chemistry: varying single-molecule junction conductance using chemical substituents. Nano Lett. 7, 502506 (2007).

104. Hammett,L.P. Effect of structure upon the reactions of organic compounds. Benzene derivatives. J.Am. Chem. Soc. 59, 96103 (1937).

105. Malen,J.A. etal. The nature of transport variations in molecular heterojunction electronics. Nano Lett. 9, 34063412 (2009).

106. Vonlanthen,D. etal. Chemically controlled conductivity: torsion-angle dependence in a single-molecule biphenyldithiol junction. Angew. Chem. Int. Ed. Engl. 48, 88868890 (2009).

107. Mishchenko,A. etal. Single-molecule junctions based on nitrile-terminated biphenyls: a promising new anchoring group. J.Am. Chem. Soc. 133, 184187 (2011).

108. Jan van der Molen,S. & Liljeroth,P. Charge transport through molecular switches. J.Phys. Condens. Matter 22, 133001 (2010).

109. Sun,L. etal. Single-molecule electronics: from chemical design to functional devices. Chem. Soc. Rev. 43, 73787411 (2014).

110. Lara-Avila,S. etal. Bianthrone in a single-molecule junction: conductance switching with a bistable molecule facilitated by image charge effects. J.Phys. Chem. C 114, 2068620695 (2010).

111. Duli,D. etal. One-way optoelectronic switching of photochromic molecules on gold. Phys. Rev. Lett. 91, 207402 (2003).

112. van der Molen,S.J. etal. Light-controlled conductance switching of ordered metalmoleculemetal devices. Nano Lett. 9, 7680 (2009).

113. Darwish,N., Aragons,A.C., Darwish,T., Ciampi,S. & Dez-Prez,I. Multi-responsive photo- and chemo-electrical single-molecule switches. Nano Lett. 14, 70647070 (2014).

114. Li,Z. etal. Single-molecule sensing of environmental pH an STM break junction and NEGF-DFT approach. Angew. Chem. Int. Ed. Engl. 53, 10981102 (2014).

115. Roldan,D. etal. Charge transport in photoswitchable dimethyldihydropyrene-type single-molecule junctions. J.Am. Chem. Soc. 135, 59745977 (2013).

116. Liljeroth,P., Repp,J. & Meyer,G. Current-induced hydrogen tautomerization and conductance switching of naphthalocyanine molecules. Science 317, 12031206 (2007).

117. Auwrter,W. etal. A surface-anchored molecular four-level conductance switch based on single proton transfer. Nat. Nanotechnol. 7, 4146 (2012).

118. Lrtscher,E., Ciszek,J.W., Tour,J. & Riel,H. Reversible and controllable switching of a single-molecule junction. Small 2, 973977 (2006).

119. Baer,R. & Neuhauser,D. Phase coherent electronics: a molecular switch based on quantum interference. J.Am. Chem. Soc. 124, 42004201 (2002).

120. Cardamone,D.M., Stafford,C.A. & Mazumdar,S. Controlling quantum transport through a single molecule. Nano Lett. 6, 24222426 (2006).

121. Hod,O., Baer,R. & Rabani,E.A. Parallel electromagnetic molecular logic gate. J.Am. Chem. Soc. 127, 16481649 (2005).

122. Vazquez,H. etal. Probing the conductance superposition law in single-molecule circuits with parallel paths. Nat. Nanotechnol. 7, 663667 (2012).

123. Gudon,C.M. etal. Observation of quantum interference in molecular charge transport. Nat. Nanotechnol. 7, 305309 (2012).

124. Rabache,V. etal. Direct observation of large quantum interference effect in anthraquinone solid-state junctions. J.Am. Chem. Soc. 135, 1021810221 (2013).

125. Darwish,N. etal. Observation of electrochemically controlled quantum interference in a single anthraquinone-based norbornylogous bridge molecule. Angew. Chem. Int. Ed. Engl. 51, 32033206 (2012).

126. Baghernejad,M. etal. Electrochemical control of single-molecule conductance by fermi-level tuning and conjugation switching. J.Am. Chem. Soc. 136, 1792217925 (2014).

127. Solomon,G.C., Herrmann,C., Hansen,T., Mujica,V. & Ratner,M.A. Exploring local currents in molecular junctions. Nat. Chem. 2, 223228 (2010).

128. Solomon,G.C. etal. Quantum interference in acyclic systems: conductance of cross-conjugated molecules. J.Am. Chem. Soc. 130, 1730117308 (2008).

129. Phelan,N.F. & Orchin,M. Cross conjugation. J.Chem. Educ. 45, 633637 (1968).

130. Klausen,R.S. etal. Evaluating atomic components in fluorene wires. Chem. Sci. 5, 15611564 (2014).

131. Yoshizawa,K. An orbital rule for electron transport in molecules. Acc. Chem. Res. 45, 16121621 (2012).

132. Tsuji,Y. & Hoffmann,R. Frontier orbital control of molecular conductance and its switching. Angew. Chem. Int. Ed. Engl. 53, 40934097 (2014).

133. Coulson,C.A. & Rushbrooke,G.S. Note on the method of molecular orbitals. Math. Proc. Cambridge Philos. Soc. 36, 193200 (1940).

134. Markussen,T., Stadler,R. & Thygesen,K.S. The relation between structure and quantum interference in single molecule junctions. Nano Lett. 10, 42604265 (2010).

135. Li,E.Y. & Marzari,N. Conductance switching and many-valued logic in porphyrin assemblies. J.Phys. Chem. Lett. 4, 30393044 (2013).

136. Qian,G., Saha,S. & Lewis,K.M. Two-state conductance in single Zn porphyrin molecular junctions. Appl. Phys. Lett. 96, 243107 (2010).

137. Li,X., Staykov,A. & Yoshizawa,K. Orbital views of the electron transport through polycyclic aromatic hydrocarbons with different molecular sizes and edge type structures. J.Phys. Chem. C 114, 999710003 (2010).

138. Yamada,R., Kumazawa,H., Noutoshi,T., Tanaka,S. & Tada,H. Electrical conductance of oligothiophene molecular wires. Nano Lett. 8, 12371240 (2008).

139. Reuter,M.G. & Hansen,T. Finding destructive interference features in molecular transport junctions. J.Chem. Phys. 141, 181103 (2014).

140. Bartels,L. Tailoring molecular layers at metal surfaces. Nat. Chem. 2, 8795 (2010).

141. Maggio,E., Solomon,G.C. & Troisi,A. Exploiting quantum interference in dye sensitized solar cells. ACS Nano 8, 409418 (2014).

142. Bergfield,J.P., Heitzer,H.M., Van Dyck,C., Marks,T.J. & Ratner,M.A. Harnessing quantum interference in molecular dielectric materials. ACS Nano 9, 64126418 (2015).

143. Landauer,R. Spatial variation of currents and fields due to localized scatterers in metallic conduction. IBM J.Res. Dev. 1, 223231 (1957).

144. Bttiker,M., Imry,Y., Landauer,R. & Pinhas,S. Generalized many-channel conductance formula with application to small rings. Phys. Rev. B 31, 62076215 (1985).

145. He,J. etal. Electronic decay constant of carotenoid polyenes from single-molecule measurements. J.Am. Chem. Soc. 127, 13841385 (2005).

146. Lu,Q. etal. From tunneling to hopping: a comprehensive investigation of charge transport mechanism in molecular junctions based on oligo(p-phenylene ethynylene)s. ACS Nano 3, 38613868 (2009).

147. Kaliginedi,V. etal. Correlations between molecular structure and single-junction conductance: a case study with oligo(phenylene-ethynylene)-type wires. J.Am. Chem. Soc. 134, 52625275 (2012).

148. Capozzi,B. etal. Length-dependent conductance of oligothiophenes. J.Am. Chem. Soc. 136, 1048610492 (2014).

149. Hines,T. etal. Transition from tunneling to hopping in single molecular junctions by measuring length and temperature dependence. J.Am. Chem. Soc. 132, 1165811664 (2010).

AcknowledgementsT.A.S. is supported by an NSF Graduate Research Fellowship under grant no. 11-44155. The authors thank the NSF for support under grant no. CHE-1404922.

Competing interests statementThe authors declare no competing interests.

R E V I E W S



Abstract | The field of single-molecule electronics harnesses expertise from engineering, physics and chemistry to realize circuit elements at the limit of miniaturization; it is a subfield of nanoelectronics in which the electronic components are single Figure 1 | A schematic of a single-molecule junction with electrode, anchor and bridge components. The bridge unit can be further deconstructed into backbone (blue block) and substituent (red circles) subunits. I, current.Anchor groupFigure 2 | Anchor group archetypes and the nature of charge carriers for common dative anchors. a|Molecular structures of common anchors. Dative anchors can be classified as donating or lone pair donating. For lone pair donors, the anchors shown in thBox 1 | Principles of charge transport in single-molecule junctionsFigure 3 | Tuning the structure of the anchor, electrode and bridge to modulate charge-transport properties in single-molecule junctions. a|Current and differential conductance plotted against source-drain voltage bias at a constant 1V gate voltage for ElectrodeMolecular bridgeTable 1 | values of oligomeric materials with conductance dominated by coherent-tunnelling mechanisms Figure 4 | Single-molecule conductance switches and quantum interference features. a|Reversible voltage-induced switching between on and off states in an oligo(phenylene ethynylene)-based wire. b|Calc

Documents

Chemical principles of single-molecule electronics - …nuckolls.chem.columbia.edu/system/files/175/original/16… · · 2017-01-29Chemical principles of single-molecule electronics